Art. 9.— T. Terada : 



luminous image is always to be traced, whicli in most cases 

 practically attains the final position after a few minutes (see PL 

 I, marked with^). This effect will of course affect the accuracy of 

 the record, especially in relatively magnifying the disturbances of 

 the longer period compared with the shorter ones. The error 

 may, indeed, amount to several percent in unfavorable cases, 

 if the relative amplitudes of the disturbances with decidedly 

 different periods are to l)e compared with accurac3^ However, the 

 general inferences which will be given in the following communi- 

 cation will not be seriously affected, since here the comparison of 

 amplitudes is either made of the X- and Y-components for waves 

 of the same period, or of the X- and Z-components for waves with 

 the periods usually longer than 1 minute. In the former case, 

 both instruments show after-effects very similar to each other. 

 Even in the latter case, where the Z-instrument is comparatively 

 free from such effect, any serious error Avill occur only in the case 

 of very short waves. This very disagrea])le time-effect could 

 prol)ably l)e avoided by the use of quartz fil)re suspension, directly 

 welded to the quartz stirrup, though in this case we must devise 

 a necessary modification of the suspension head. 



h) North-South- or X-Component Instrument (North taken as 

 positive). The NS- insti'ument is essentially the same as that of 

 Mascart's magnetograph. The suspension system is, however, 

 replaced by one quite similar to that of the WE- instrument above 

 described, except that in the present case the plane of the mirror 

 is perpendicular to the axis of the magnetic needle. 



Keeping the breadth of the l)i filar suspension at first suffici- 

 ently wide, the torsion head is slowly twisted, till the luminous 

 spot appears in the assigned range. Then cautiously turn the 

 torsion head, at the same time regulating the breadth of the bifilar 

 suspension, so as to keep the luminous image always within 

 proper range. The unstable position is attained in a certain 

 azimuth of the torsion head, which may be noted down for 

 subsequent readjustment. The determination of the constant is 

 carried out in a way similar to that in the case of the Y-instrument, 

 with the only difference that in this case the defiecting magnet is 



